An artificial neural network based approach to fault diagnosis and classification of fluid power systems

Abstract

In this paper, multilayer perceptron (MLP) type neural networks are used to detect leakages in an electrohydraulic cylinder drive. Both single-leakage and multiple-leakage type faults are investigated. The performance of MLPs is examined relating to the level of leakage flowrate and it was found that MLPs perform well for line leakages but for across-cylinder seal leakages they could only detect leakage over 1.01/min. The generalization tests on non-training leakage flowrate and working temperature are also included. A novel feature is the use of system state variables for network training, including additional terms to accelerate convergence. The approach has also made a significant contribution to multiple-fault detection, particularly for the complex three-fault case.